BOSTON — Tucked away in a small room down a hallway from the main festivities of the MIT Sloan Sports Analytics Conference last weekend, a group of luminaries in the sports data revolution gathered for a makeshift curling tournament.

There was Daryl Morey, general manager of the Houston Rockets; Nate Silver, creator of the website FiveThirtyEight; and John Hollinger, formerly with ESPN and now a member of the Memphis Grizzlies’ front office. Also in the bracket of the double-elimination, 12-person field was Jessica Gelman, an executive with the Kraft Analytics Group who co-founded the conference, and Mike Zarren, the Boston Celtics’ assistant general manager and salary cap expert.

As the amateur curlers tossed their stones — discussions about bringing in a sheet of ice were had, but in the end the games were played without brooms and on large plastic boards – two gurus from the world of curling analytics, Gerry Geurts and Kevin Palmer, looked on and offered advice.

“I love curling; it’s like chess on ice,” Morey said. “All analytics really is is analyzing what wins. You start with win probability and work backwards. This isn’t as complex as basketball, but there’s real work to do, and these guys know what they’re talking about.”

A day before, Geurts and Palmer presented a research paper, “The Evolution of Curling Analytics,” making the convergence of the sport and the conference more than just high jinks.

Geurts and Palmer consult for several national teams about implementing statistical analysis into their curling strategies. One of their clients was the gold medal-winning United States men’s team at the 2018 Pyeongchang Olympics.

“The analytics has been revolutionary for the sport,” said Tyler George, a member of that gold-medal winning team who attended the conference (and won the curling tournament).

Curling also represents another frontier conquered by data.

“There are panels here this weekend about chess and poker,” Silver said. “So it’s broadening the definition of analytics and sports – and also the overall geekiness of the conference.”

Geurts grew up near London, Ontario. He curled but was also a huge baseball fan, listening to Toronto Blue Jays games on the radio and keeping score with a system he devised himself.

He was already working to digitize curling data through his company, CurlingZone, and make it more widely available when “Moneyball,” the seminal Michael Lewis book about analytics and the Oakland Athletics, was published in 2004. Geurts believed that his sport could make advances in analysis, too.

“It’s the perfect sport for it because of the precise moments of decision-making,” he said. “You have to figure out how to deal with each stone, and there’s a very specific point of strategy.”

A quick refresher on curling: A game lasts for 10 ends, which are like baseball’s innings, in which each team throws eight rocks toward a bull’s eye target called the button. The team with the closest rock to the center scores.

By the mid-2000s, Geurts hooked up with Palmer, who grew up in Winnipeg and was a voracious reader of Bill James, the godfather of the analytics movement in baseball. The two began crunching numbers, and among their findings was a startling discovery that contradicted conventional wisdom in the sport: It is better to be ahead by a point in an end with the other team throwing the hammer (the last stone) than to be trailing by a point and have the hammer.

In other words, it was usually better to have less control over the final shot.

“Overcoming that human element of wanting control was a hard thing,” Palmer said.

Despite more than 10-plus years of study, Geurts and Palmer have found the adoption of their research slow, and not until the past few years have the two made much traction. They met a group of American teams, including George’s, at the U.S. Olympic headquarters in Colorado Springs in the summer of 2017 and presented their research as well as analyses of top teams around the world.

“You could see there were weaknesses in these great teams,” George said. “That’s really helpful when you’re an underdog.”

Geurts also worked extensively with the women’s gold-medal winning Swedish team in Pyeongchang, helping propel them to an 8-3 victory over South Korea in the final. He was also working for the Canadian Broadcasting Corporation during the Olympics and texted an announcer ahead of time that he was sorry the championship game was going to be so boring. “The South Koreans had an aggressive style that no one matched, so it was easy to plan for,” he said.

Geurts now consults for about 20 teams around the world. Moving forward, Geurts, Palmer and a third partner, Jason Gunnlaugson, are hoping to dig deeper into data. There are Canadian researchers putting sensors on stones that will track speed and spin rate and help gather more types of data for them to process.

The presentation of their paper wasn’t the most well-attended event of the weekend, but there were dozens in the audience. One student pursuing a PhD at MIT, Nathaniel Bailey, approached Geurts afterward for job advice.

“I’m looking for a career in analytics, and I’m on the curling team here,” he said. “But I know in curling it’s a difficult career path.”

Copy the Story Link